Training annotations and octopi predictions of ribosomes in unroofed human cells

Deposition Overview

This deposition provides manual and automated annotations of cytosolic ribosomes in cryo-electron tomograms of unroofed human cells. Manual annotations were created using the copick-toolkit and used to train an octopi object detection model. The trained model was used to localize candidate particles, which were then subjected to subtomogram averaging and classification using RELION 5.0. The final set of automated annotations is the set of particles that contributed to the highest-quality ribosome average (7.5 Å final resolution).